search

weaviate / i2v-pytorch-models

Inference containers for the Weaviate `img2vec-pytorch` module
BSD 3-Clause "New" or "Revised" License
1 stars 4 forks source link

Issue when importing img #1

Open bobvanluijt opened 3 years ago

bobvanluijt commented 3 years ago

When importing images. The i2vec container outputs: OpenBLAS Warning : Detect OpenMP Loop and this application may hang. Please rebuild the library with USE_OPENMP=1 option.

etiennedi commented 3 years ago

Thanks, is this only a warning or does this prevent importing of some sorts? cc @antas-marcin

bobvanluijt commented 3 years ago

It prevents importing. Disclaimer, I'm using my new M1 Mac, I don't know if there is a relation

etiennedi commented 3 years ago

I still have the M1 Mac mini here, so I can take a quick look. Do you have a repo I can clone that imports some images?

bobvanluijt commented 3 years ago

Yes: https://github.com/semi-technologies/DEMO-magento-e-commerce

On Fri, Jul 23, 2021 at 12:30 PM Etienne Dilocker @.***> wrote:

I still have the M1 Mac mini here, so I can take a quick look. Do you have a repo I can clone that imports some images?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/semi-technologies/i2v-pytorch-models/issues/1#issuecomment-885547745, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABKBAKSGZF3CDRDTXMJXT6LTZFAFBANCNFSM5A3UZKQQ .

-- SeMI Technologies website https://www.semi.technology - Twitter https://twitter.com/SeMI_tech - LinkedIN https://www.linkedin.com/company/semi-technologies - P.O. Box 95263, 1090HG Amsterdam

etiennedi commented 3 years ago

I'm seeing this on the M1 Mac:

Cross referencing product with product ----- 707 CrossSell
Cross referencing product with product ----- 481 Related
Cross referencing product with product ----- 325 UpSell
Cross referencing product reviews ---------- 1000
Cross referencing product wishlists -------- 14
Importing images into Weaviate ------------- 9

With the last number going up one at a time. Does that mean I'm not running into this issue or is it still occurring, but just hidden somewhere?

EDIT:

I think it completed successfully, it printed:

Importing images into Weaviate ------------- 462
Cross referencing product images ----------- 4088
Total time required ------------------------: 5 minutes 36.5 seconds
bobvanluijt commented 3 years ago

The output is showing in the logs of the Docker container for me. Please note that also images are not added

On Fri, 23 Jul 2021 at 13:15 Etienne Dilocker @.***> wrote:

I'm seeing this on the M1 Mac:

Cross referencing product with product ----- 707 CrossSell Cross referencing product with product ----- 481 Related Cross referencing product with product ----- 325 UpSell Cross referencing product reviews ---------- 1000 Cross referencing product wishlists -------- 14 Importing images into Weaviate ------------- 9

With the last number going up one at a time. Does that mean I'm not running into this issue or is it still occurring, but just hidden somewhere?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/semi-technologies/i2v-pytorch-models/issues/1#issuecomment-885568832, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABKBAKRPOZGHN525CIMZDGTTZFFMTANCNFSM5A3UZKQQ .

-- SeMI Technologies website https://www.semi.technology - Twitter https://twitter.com/SeMI_tech - LinkedIN https://www.linkedin.com/company/semi-technologies - P.O. Box 95263, 1090HG Amsterdam

etiennedi commented 3 years ago

I can also see the warnings, but both the images as well as their vectors look fine for me:

Screenshot 2021-07-23 at 13 25 45 Screenshot 2021-07-23 at 13 25 26

bobvanluijt commented 3 years ago

I can also see the warnings, but both the images as well as their vectors look fine for me

Is this on the M1 machine?

etiennedi commented 3 years ago

Yes :)